AIQ AIQ
Learning from Examples · Lesson 3.1.1

Teaching "Learning from Examples" to Explorer mode (ages 5–7)

Part of the Learning from Examples lesson guide. Teaching a different grade? 🔧 Builder (8–10) · 💻 Hacker (11–14) · ⚡ Architect (15–18)

Hook & Warm-Up

Gather the class on the rug or have everyone turn to face you — this works as call-and-response, not silent listening. Hold up your hands like a big question mark and read the app's own opening line with real energy, as Pixel (the Explorer mascot):

"How did you learn what a dog looks like? You saw LOTS of dogs! 🐕
A.I. learns the same way — by looking at tons of examples!"

Ask right away: "Who has a dog, or an aso, at home — or maybe your lolo's neighbor has one? How did you know it was a dog and not, say, a cat, the very first time someone told you?" Take two or three answers, and accept anything a child offers ("it has four legs," "it barks," "my mama told me"). Land on the idea gently: "That's it exactly — someone SHOWED you dogs, again and again, and told you the name. Today we're going to see a smart helper learn a new thing the exact same way you did."

Point to the four little pictures the app shows at the start of this lesson — a cat, a little tag (that's a "label"), a stack of books (that's "examples"), and a target (that's "guess!"). Tell the class: "Keep an eye on these four friends today: a cat, a name-tag, lots of examples, and a guess."

Optional warm-up game before you even touch pictures: ask three or four students, one at a time, "What is your favorite animal, and how did you first learn its name?" Almost every answer will involve someone showing them the animal (in real life, in a book, or on a screen) more than once. Point out: "Notice nobody just told you a rule like 'four legs plus fur equals dog' — you just SAW dogs, again and again, until you knew. That's the whole secret we're exploring today."

Main Activity

Before opening the app, do a two-minute physical warm-up so the idea is in their hands before it's on a screen. Bring — or quickly draw on the board — five or six pictures of cats and five or six of dogs (aso and pusa work well, or use whatever animals are easy to find). Show each one and say its name out loud as you show it: "Cat!" "Dog!" "Cat!" Then hold up a brand-new picture the class hasn't seen and ask: "Now YOU guess — cat, or dog?" Let a few kids guess out loud. Whatever they say, respond warmly: if they're right, "Yes! You learned it!"; if they're not quite there yet, "Ooh, close — let's look again, no worries, guessing is how we learn" (never say "wrong" — the app itself never does). This little game IS supervised learning, just done with your own two hands instead of a computer.

Now open the lesson on the shared screen. It walks through the same idea in three parts, and your job is to translate each screen into words this age already understands:

🏷️ Supervised Learning — The app shows a labeled photo pile. Say: "Grown-ups helped teach the smart helper by showing it LOTS of pictures and telling it the name of each one — just like your mama or teacher taught you." Then: "The more pictures it sees, the smarter it gets — just like how the more times you see a dog, the faster you know it's a dog." Then it shows the smart helper being given a brand-new picture it's never seen, and guessing. Say: "Now comes the fun part — we hide the answer and see if it can guess all by itself! Sometimes it gets it right away, and sometimes it needs a little more practice — that's totally okay."

👨‍🏫 Teaching AI — "When the smart helper guesses right, it says to itself, 'yes, that worked, I'll remember to do that again!' And when it's not quite right yet, it doesn't get sad or give up — it just changes its thinking a teensy bit and tries again." Explain "practicing the same pictures again and again" by comparing it to something the class actually does: "It's like practicing your ABCs, or your ten jumping jacks — you don't just do it once, you do it lots of times until it sticks." Mention gently that really big smart helpers can take many, many days of practicing, using lots of computers working together — "way more practicing than even the longest school year!"

🌍 Real Examples — Go through each one slowly and connect it to something the class has actually seen, pausing after each for a thumbs-up if anyone recognizes it:

Ask the class after this list: "Do you notice something all four of these have in common? Someone had to show them a GIANT pile of already-answered examples before they could guess anything at all." That one sentence is the whole lesson, said back to them in their own moment of noticing it.

Let students finish the app's short matching round on their own or in pairs — it pairs things that "had to learn from examples" (a labeled photo, a trained spam filter, a song-guesser, a weather guesser) against everyday things that never need to learn anything at all (a paper clip, a mailbox, a pair of drumsticks, a glass thermometer). Frame it simply as: "Which ones had to practice with lots of examples, and which ones just... are what they are, no learning needed?"

As students finish the matching round, circulate and ask one or two, quietly, "How could you tell which side of each pair had to learn?" Listen for anything close to "because it has to guess" or "because it gets smarter the more it sees" — that's the real idea, in their own words, and it's worth repeating back to them so they hear it confirmed: "Exactly — a paper clip is always just a paper clip, no matter how many times you use it. But a song-guesser gets better the more songs it hears."

Discussion

Quiz Walkthrough

Read each question aloud before students answer on their own device, and if a hand goes up unsure, retell the cat-and-dog warm-up in one sentence rather than just repeating the question — it's the memory hook this whole quiz is testing.

A.I. learns from... (📚 Lots and lots of examples! / Electricity / Some kind of magic process / Just one picture)
📚 Lots and lots of examples! There's no magic and no need for just electricity to run on — a smart helper gets better the exact same way the class did with the cat-and-dog pictures: by seeing many, many examples with the answer attached.
When A.I. guesses wrong, it... (Gives up / Gets mad / Goes to sleep / Tries again a little differently!)
Tries again a little differently! A not-quite-right guess is never the end — it's exactly how a smart helper keeps learning, the same way the class kept guessing during the warm-up game until they got it.
What makes A.I. better at learning? (No examples / Less examples / Being louder / More examples!)
More examples! The more pictures, sounds, or pieces of data a smart helper sees, the more chances it has to notice what makes a cat a cat, or a dog a dog.
Can A.I. learn from just ONE picture? (Yes! / Not really — it needs lots! / Only small ones / Only big pictures primarily)
Not really — it needs lots! One example isn't enough to notice a real pattern — that's true for the class too: seeing one dog one time wouldn't have been enough to know what "dog" means for every dog in the world.

Wrap-Up & Extension

Close with: "Today you and a smart helper both learned the exact same way — by looking at LOTS of examples, guessing, and trying again when we weren't quite there yet. From now on, whenever you see something new, remember: the more examples you see, the better you get at guessing!"

Extension activity — "Guess the Fruit" card game: Using real fruit, drawings, or magazine cutouts, build two small piles of Filipino fruits — say, mangga and saging. Show five of each, naming them as you go, exactly like the cat-and-dog warm-up. Then reveal a brand-new picture and let the class guess. Repeat with a bigger pile — ten of each instead of five — and ask if the guessing got easier. This makes the "more examples = better guesses" idea concrete and stretches the 10–15 minute core lesson into a full 30–40 minute activity. For an extra round, add a third fruit (papaya) partway through with only one or two examples, and let the class notice out loud that it's much harder to guess correctly with so few pictures to learn from. If your class has extra time, try a second round with something less visual — clap a simple rhythm and have students copy it, then vary it slightly and see if they can now tell "clap patterns that match" from "clap patterns that don't," to show that learning-from-examples isn't only for pictures. End the extension by asking the whole class together: "So — do you need lots of examples, or just one, to really learn something new?" and let them answer in a happy chorus, since this is exactly the quiz question they'll see again.

← Lesson overview ← AI Perception Challenge (Explorer) Finding Patterns (Explorer) →